Dysfunction in GABA signalling mediates autism-like stereotypies and Rett syndrome phenotypes

Abstract

Mutations in the X-linked MECP2 gene, which encodes the transcriptional regulator methyl-CpG-binding protein 2 (MeCP2), cause Rett syndrome and several neurodevelopmental disorders including cognitive disorders, autism, juvenile-onset schizophrenia and encephalopathy with early lethality. Rett syndrome is characterized by apparently normal early development followed by regression, motor abnormalities, seizures and features of autism, especially stereotyped behaviours. The mechanisms mediating these features are poorly understood. Here we show that mice lacking Mecp2 from GABA (γ-aminobutyric acid)-releasing neurons recapitulate numerous Rett syndrome and autistic features, including repetitive behaviours. Loss of MeCP2 from a subset of forebrain GABAergic neurons also recapitulates many features of Rett syndrome. MeCP2-deficient GABAergic neurons show reduced inhibitory quantal size, consistent with a presynaptic reduction in glutamic acid decarboxylase 1 (Gad1) and glutamic acid decarboxylase 2 (Gad2) levels, and GABA immunoreactivity. These data demonstrate that MeCP2 is critical for normal function of GABA-releasing neurons and that subtle dysfunction of GABAergic neurons contributes to numerous neuropsychiatric phenotypes.

Figure 1. Viaat-Mecp2^−/y mice lose MeCP2 in GABA⁺ neurons and develop stereotypies, self-injury, and compulsive behavior

(a) WT cortex layer 2/3 neurons from 17-week old mice labeled with DAPI, MeCP2, and GABA reveal 50% higher MeCP2 levels in GABA⁺ (circled) than in GABA⁻ cells (asterisk). Data normalized to MeCP2 level in GABA⁻ cells; n = 3 mice per genotype. (b) Viaat-Cre expression as assessed by ROSA26R-EYFP reporter and colocalization of EYFP and GABA in 14-week old mice. (c) More than 90% of GABA⁺ cells in Viaat-Mecp2^−/y mice lack MeCP2. Data from n = 3 mice per genotype. (d) Seven-week old Viaat-Mecp2^−/y mice displaying forepaw and hindlimb clasping (arrowhead). (e) Viaat-Mecp2^−/y mice showing fur loss at 15 weeks of age and self-injury, including ocular damage, at 24 weeks (arrowhead). (f,g) Viaat-Mecp2^−/y mice show ~300% increase in grooming time (f) and in number of holes explored with ≥ 2 sequential nose-pokes (g).

Figure 2. MeCP2 deficiency in GABAergic neurons causes several RTT-like features

(a-f) Viaat-Mecp2^−/y mice have more footslips (a), reduced number of side touches on dowel (b), shorter latency to fall on rotarod (c) and wire (d), reduced forelimb grip strength (e), and pronounced hypoactivity (f). (g) Viaat-Mecp2^−/y mice have intact social recognition but increased social interaction with novel and familiar partners. (h) Viaat-Mecp2^−/y mice spend 60% more time interacting with the unfamiliar stranger mouse than controls. The wire cup served as a familiar inanimate control without social valence. (i) Viaat-Mecp2^−/y mice exhibit similar interaction time with a novel inanimate Lego™ object compared to controls. (j) Viaat-Mecp2^−/y mice are poor nest-builders. (k,l) Viaat-Mecp2^−/y mice show impaired maximum ASR to 120 dB (k) and increased PPI at 78 and 82 dB prepulses (l). (m,n) Viaat-Mecp2^−/y mice have similar learning rate during training (m) but reduced crossings over the target platform location during the probe test (n).

Figure 3. Loss of MeCP2 in inhibitory GABAergic neurons compromises respiration and survival

(a) Viaat-Mecp2^−/y mice exhibit premature lethality with 50% survival by 26-weeks. Dlx5/6-Mecp2^−/y mice survive up to at least 80-weeks. (b) Representative plethysmography traces from 32-week old WT, Flox, Dlx5/6-Cre, Viaat-Cre, Dlx5/6-Mecp2^−/y, and Viaat-Mecp2^−/y mice. Only Viaat-Mecp2^−/y mice exhibit pronounced apneas and abnormal respiratory pattern. (c-e) Viaat-Mecp2^−/y mice show 42% reduction in tidal volume (c), 45% reduction in minute volume (d), and increased number of apneas longer than 0.4 s (e). (f-h) Dlx5/6-Mecp2^−/y mice show no alterations in tidal volume, minute volume, and no apneas.

Figure 4. MeCP2 deficiency in GABAergic neurons reduces Gad1, Gad2, and GABA levels

(a, b) Somatic GABA immunoreactivity in GABA⁺ cells from 15-17-week old Flox and Viaat-Mecp2^−/y mice labeled with DAPI, MeCP2, and GABA. GABA⁻ cells are labeled with asterisks. Images show loss of MeCP2 and reduced GABA immunoreactivity in GABA⁺ cells (circled) of Viaat-Mecp2^−/y mice by 37% in cortical layer 2/3 neurons (a) and 50% in striatal neurons (b). Data normalized to GABA level in WT; n = 2-4 mice per genotype. (c, d) Gad1 and Gad2 mRNA levels are reduced in Viaat-Mecp2^−/y cortex by 36% and 28%, respectively (c), and in Viaat-Mecp2^−/y striatum by 54% and 62%, respectively (d). (e) Gad1 and Gad2 mRNA levels are unaltered in CamKIIα-Mecp2^−/y cortex. (f) ChIP reveals MeCP2 occupancy of Gad1 and Gad2 promoters in WT, which is absent in IgG and KO. (g, h) Mapping MeCP2 occupancy upstream of TSS, after normalization with IgG, reveals increased occupancy in WT (black line) across Gad1 (g) and Gad2 (h) promoters without enhanced MeCP2 binding in KO (red line).

Figure 5. MeCP2 deficiency in GABAergic neurons results in reduced mIPSC quantal size in cortical layer 2/3 and striatal neurons, EEG hyperexcitability, and impaired hippocampal LTP

(a-d) Data from three mice per genotype and the number of neurons recorded is shown. (a, b) mIPSC amplitude and charge are reduced in Viaat-Mecp2^−/y cortical slices. Average traces for each genotype are overlaid and graphs show mIPSC amplitude and charge (a), and frequency (b). (c, d) mEPSC amplitude and charge are unaltered in Viaat-Mecp2^−/y cortical slices. Average traces for each genotype are overlaid and graphs show mEPSC amplitude and charge (c), and frequency (d). (e-h) Data from two independent autaptic striatal cultures and the number of neurons recorded is shown. Average traces for each genotype are overlaid and bar graphs show KO neurons have reduced mIPSC amplitude and charge (e), but no differences in frequency (f). (g) 5 μM GABA evokes similarly sized responses from WT and KO. (h) PPR is similar between WT and KO. (i) EEG recordings from constitutive null and Viaat-Mecp2^−/y mice compared to WT. Constitutive null Mecp2^−/y mice (n = 4) occasionally develop electrographic seizures, but predominantly exhibit hyperexcitability discharges. Viaat-Mecp2^−/y mice (n = 7) frequently exhibit hyperexcitability discharges, but do not show electrographic seizures. (j-l) Acute hippocampal slices from 11-13-week old mice, six mice per genotype, reveal reduced magnitude of TBS-LTP (j) and saturating LTP shows no further increases in synaptic potentiation (k) in Viaat-Mecp2^−/y slices. Number of slices recorded shown in the figure. (l) Significant increase in potentiation with the second TBS in controls but not in Viaat-Mecp2^−/y slices.